Pancreatic cancer is the fifth leading cause of cancer death in this country and it is almost uniformly fatal. The poor prognosis of pancreatic cancer patients can be attributed to the lack of early detections and effective treatments. Efforts to fight pancreatic cancer have been hampered by the absence of a good animal model of pancreatic cancer. A reliable mouse model is urgently needed. Tremendous progress has been made in understanding the molecular genetics of human pancreatic adenocarcinoma within the last decade. It is now clear that pancreatic cancer is a genetic disease; the tumor-suppressor genes most frequently inactivated include pl6/INK4a and SMAD4/DPC4, while oncogene HER-2/neu/ERBB2 is often overexpressed in human pancreatic adenocarcinoma. This valuable information can be utilized to create mouse models that directly mirror human pancreatic adenocarcinoma. The main objective of the project is to create mouse models of ductal pancreatic adenocarcinoma. Here we propose three complementary strategies to accomplish this goal. The first aim is to create a conditionally deleted p16 mouse. The second aim is to characterize SMAD4/MT-TGFalpha mice. Our third aim targets HER-2. HER-2 is upregulated at early stage of pancreatic tumorigenesis. We reason that pancreas-specific expression of HER-2 in a transgenic mouse line would provide the necessary first-hit to complement our p16 and SMAD4 knock-out mice. Our strategies emphasize the uses of conditional gene targeting, which would allow pancreas-specific gene regulation and tumor development, and avoid the development of undesirable phenotypes in mice. The urgency to detect and treat this deadly disease has prompted the NCI to issue a Progress Review Group report, in which the NCI identifies the development of an animal model of pancreatic cancer as one of its research priorities. This proposal is 100% pancreatic cancer relevant.